The second beta version of Qt 3.1 is now released. It can be downloaded from here. There is also a changelog. Additionally, Qt# 0.5 was released a few days ago, same time as Mono 0.16.
3.1b2 is in KDE’s CVS (qt-copy module) since a week or so. The new beta fixed some bugs, introduced a few new ones, nothing spectacular. The biggest advancement of 3.1 is probably in Qt Designer which is a huge improvement over 3.0. And, not to forget, Qt 3.1 finally seems to solve my headaches caused by their multi-threading support.
I notice Mono is quite biased against Qt#, notice http://go-mono.com/ has GTK# listed….. and not QT#. Never used neither, have no idea how good is either 🙂 But what amazes me is that Mono lists an outside project and not another outside project too…..
Well, *big* suprise, coming from the people who supported GTK+ and made GTK+ what it is today….
I’m asking anybody who’s using QT 3.1 beta2 out there: is it worth the update yet?
I’m curious about it, as I am writing some fairly complex multithreaded code using the QThread/etc API; it sounds like the new Qt 3.1 has better threading, so I thought I might give it a shot.
On the other hand, I heard about all sorts of binary compatibility breakages from 3.1 beta1. Though according to the changelog those troubles are fixed.
So, anybody out there who’s used it, particularly those out there who’ve used it for multithreaded programming: is it worth the switch yet?
I’ll probably just wait for the official release anyhow, as my system is stable and I don’t particularly feel like tempting fate
I think the Mono team is in for a big shock. I expect that Microsoft is going to target Mono with patent violation as soon as the project starts to “take off”. It will be unfortunate, but I recall everyone warning of it some time ago. I read the other day that Microsoft has spent millions developing .NET, and they will protect their patents when questioned. They also made public that the FreeBSD CLI was for educational purposes only. I wish I could still find the article.
I wasn’t spreading FUD, I was simply passing on my opinion. Microsoft holds several patents and Ballmer made it clear in the article that he would enforce them. I’ll see if I can find it.
OS X does flicker free resizing at the cost of inordinate amounts of memory (for the window buffers). I don’t think that’s worth it, not now, not for a long time in the future. Besides, I don’t really see the problem. I’m extremely sensitive to these things, and I cannot detect any flicker, either with Qt 3.1 beta2 or Qt 3.0.5. Most KDE applications don’t flicker either. I can whip Konqueror and KWord around as quick as I want and the only thing I ever get is slight rubber banding in Konqueror (excusable given the tremendous complexity of resizing/re-laying-out a webpage, which not even IE can do any faster).
BTW> I wouldn’t upgrade to 3.1 beta2 just yet. I’ve been having this problem with Konqueror segfaulting whenever I close it. I’m using KDE 3.1 beta2, I’d presume 3.0.x would also be affected.
Regarding the possibility of mono getting sued by Microsoft: if you have billions in cash to buy lawyer time with, you don’t have to have a solid case — you can simply harrass the mono project with groundless-but-plausible lawsuits until they run out of money for legal fees, at which point you win by default. So even everything mono does is perfectly legal, Microsoft can still shut them down whenever they feel like it.
Regarding using massive amounts of RAM to avoid flickery GUIs: as RAM becomes ever cheaper, “wasting” it on things like this starts to make more sense. In a couple of years, when new machines come with 4GB RAM in the base model, wasting a few hundred megs on bitmap caches will be seen as negligible overhead.
I’m not so much worried about the RAM overhead as other factors. First, its architecturally dirty. With a *vector* based GUI, one would ideally store a (compact!) *vector* representation rather than a bitmap representation. Second, it doesn’t scale. As you get into higher resolutions, the memory consumed by bitmaps gets huge. Vector GUIs are the future for a reason. By the time we have 4GB of RAM in a base system, I really hope we have some nice 300 dpi screens too Third, we’re memory bandwidth bound to begin with (and will be for the forseeable future). The more data you have in memory the more you hit the (slow) memory subsystem and the more you blow your caches. With processor speeds ratcheting up faster than either memory sizes or speeds, it makes far more sense to trade CPU time (re-rendering damaged areas) for memory space/bandwidth (storing undamaged contents). I’m convinced the only reason Apple went with the bitmap approach is because it makes it easy to do the useless window transparency effect. Trading off real performance for a showy gimick? Sounds like Apple to me
Mere speculations, my friend. I read some articles that Microsoft do hold patents regarding .NET Framework, but ABSOLUTELY none of them ever mentioned the names of the patents, or their number.
So I’ll take this as a plain anti-Microsoft bashing. Besides, Microsoft couldn’t possibly dare to sue Mono because of the DOJ hot on their heels and every move they make for the next 10-15 years or so would be watched closely.
Besides, remember the Samba and patents issue? Microsoft couldn’t stop them, right? The most I would expect is Microsoft forbiding them to use copyleft license (which mono don’t use is some parts…)
Rayiner, Apple may be the first commerciallzing it, but if the rumours are true about Longhorn/Blackcomb, Microsoft would be joining ranks with Apple. Maybe only then would some Linux companies invest in Fresco…
(Where oh where is my little NeWS, oh where oh where could it be?……..)
3.1b2 is in KDE’s CVS (qt-copy module) since a week or so. The new beta fixed some bugs, introduced a few new ones, nothing spectacular. The biggest advancement of 3.1 is probably in Qt Designer which is a huge improvement over 3.0. And, not to forget, Qt 3.1 finally seems to solve my headaches caused by their multi-threading support.
I notice Mono is quite biased against Qt#, notice http://go-mono.com/ has GTK# listed….. and not QT#. Never used neither, have no idea how good is either 🙂 But what amazes me is that Mono lists an outside project and not another outside project too…..
Well, *big* suprise, coming from the people who supported GTK+ and made GTK+ what it is today….
The site has Qt# listed, in Resources.
Gtk# is less of an outside project than Qt#, Gtk# is in the mono cvs repository for example.
But still, browsing through the archives of the mailinglists, the people of the mono project seem pretty supportive for the Qt# project.
I’m asking anybody who’s using QT 3.1 beta2 out there: is it worth the update yet?
I’m curious about it, as I am writing some fairly complex multithreaded code using the QThread/etc API; it sounds like the new Qt 3.1 has better threading, so I thought I might give it a shot.
On the other hand, I heard about all sorts of binary compatibility breakages from 3.1 beta1. Though according to the changelog those troubles are fixed.
So, anybody out there who’s used it, particularly those out there who’ve used it for multithreaded programming: is it worth the switch yet?
I’ll probably just wait for the official release anyhow, as my system is stable and I don’t particularly feel like tempting fate
I think the Mono team is in for a big shock. I expect that Microsoft is going to target Mono with patent violation as soon as the project starts to “take off”. It will be unfortunate, but I recall everyone warning of it some time ago. I read the other day that Microsoft has spent millions developing .NET, and they will protect their patents when questioned. They also made public that the FreeBSD CLI was for educational purposes only. I wish I could still find the article.
Aah, the typical FUD about Mono. Don’t develop for it! It will die the moment it’s any good!
All the basic functional abilities of .Net can not be patented, because, well, it’s not innovative enough 🙂 Prior art exists for it.
Anyway, there’s a faq about it on the go-mono.org site..
I wasn’t spreading FUD, I was simply passing on my opinion. Microsoft holds several patents and Ballmer made it clear in the article that he would enforce them. I’ll see if I can find it.
WHEN Qt will have flicker-free resizing?
GTK+ added this for version 2.0. GTK+ implementation is a very simple one, it is not as good as the MacOSX one, but still, better than nothing.
Qt still does not really supports it by default! All my KDE apps redraw like hell when I resize them. >:(
Here is a summary, I don’t know why I can’t find the actual article. If anyone else has read it, and saved the link, please post it.
http://swpat.ffii.org/players/microsoft/index.en.html
OS X does flicker free resizing at the cost of inordinate amounts of memory (for the window buffers). I don’t think that’s worth it, not now, not for a long time in the future. Besides, I don’t really see the problem. I’m extremely sensitive to these things, and I cannot detect any flicker, either with Qt 3.1 beta2 or Qt 3.0.5. Most KDE applications don’t flicker either. I can whip Konqueror and KWord around as quick as I want and the only thing I ever get is slight rubber banding in Konqueror (excusable given the tremendous complexity of resizing/re-laying-out a webpage, which not even IE can do any faster).
BTW> I wouldn’t upgrade to 3.1 beta2 just yet. I’ve been having this problem with Konqueror segfaulting whenever I close it. I’m using KDE 3.1 beta2, I’d presume 3.0.x would also be affected.
> I don’t think that’s worth it
I think it does. Tastes…
Regarding the possibility of mono getting sued by Microsoft: if you have billions in cash to buy lawyer time with, you don’t have to have a solid case — you can simply harrass the mono project with groundless-but-plausible lawsuits until they run out of money for legal fees, at which point you win by default. So even everything mono does is perfectly legal, Microsoft can still shut them down whenever they feel like it.
Regarding using massive amounts of RAM to avoid flickery GUIs: as RAM becomes ever cheaper, “wasting” it on things like this starts to make more sense. In a couple of years, when new machines come with 4GB RAM in the base model, wasting a few hundred megs on bitmap caches will be seen as negligible overhead.
I’m not so much worried about the RAM overhead as other factors. First, its architecturally dirty. With a *vector* based GUI, one would ideally store a (compact!) *vector* representation rather than a bitmap representation. Second, it doesn’t scale. As you get into higher resolutions, the memory consumed by bitmaps gets huge. Vector GUIs are the future for a reason. By the time we have 4GB of RAM in a base system, I really hope we have some nice 300 dpi screens too Third, we’re memory bandwidth bound to begin with (and will be for the forseeable future). The more data you have in memory the more you hit the (slow) memory subsystem and the more you blow your caches. With processor speeds ratcheting up faster than either memory sizes or speeds, it makes far more sense to trade CPU time (re-rendering damaged areas) for memory space/bandwidth (storing undamaged contents). I’m convinced the only reason Apple went with the bitmap approach is because it makes it easy to do the useless window transparency effect. Trading off real performance for a showy gimick? Sounds like Apple to me
Mere speculations, my friend. I read some articles that Microsoft do hold patents regarding .NET Framework, but ABSOLUTELY none of them ever mentioned the names of the patents, or their number.
So I’ll take this as a plain anti-Microsoft bashing. Besides, Microsoft couldn’t possibly dare to sue Mono because of the DOJ hot on their heels and every move they make for the next 10-15 years or so would be watched closely.
Besides, remember the Samba and patents issue? Microsoft couldn’t stop them, right? The most I would expect is Microsoft forbiding them to use copyleft license (which mono don’t use is some parts…)
Rayiner, Apple may be the first commerciallzing it, but if the rumours are true about Longhorn/Blackcomb, Microsoft would be joining ranks with Apple. Maybe only then would some Linux companies invest in Fresco…
(Where oh where is my little NeWS, oh where oh where could it be?……..)